ferrocenic acid derivatives: towards rationalizing changes in the electronic and geometric...

Journal of Organometallic Chemistry 556 (1998) 11–20

Ferrocenic acid derivatives: towards rationalizing changes in theelectronic and geometric structures

Lily Lin, Attila Berces, Heinz-Bernhard Kraatz *

NRC-Steacie Institute for Molecular Sciences, 100 Sussex Dri6e, Ottawa, ON K1A 0R6, Canada

Received 8 July 1997; received in revised form 24 November 1997

Abstract

Using calculations based on Density functional theory (DFT), we optimized the geometry of three compounds containing theferrocenoyl moiety, ferrocene aldehyde (I), ferrocene carboxylic acid (II), ferrocene amide (III) and the parent ferrocene. Thecalculated results were compared to experimental structural data. General trends have emerged: the carbonyl C–O distanceincreases in the order aldehydeBketonesBamidesBacid. The O–C–X (X=O, N, C) angles are less responsive to substituentsand seem to be more a reflection of packing forces. Expectedly the distance between carbonyl carbon and the organic group R(R�OH (acid), NR2 (amide), CR3 (ketone) increases in the order acidBamidesBketones. Maximization of electron delocalizationinvolving the C(O)–X (X=O, N, C) and the Cp ring will force the Cp ring and the C(O)–X substituent to be co-planar.However, steric interference between the Cp ring and the substituent can result in a significant deviation from co-planarity andcan give rise to a significant twist. © 1998 Elsevier Science S.A. All rights reserved.

Keywords: Ferrocene; Ammonium ion; Electronic and geometric structures

1. Introduction

In view of the increasing importance of ferrocenes inasymmetric catalysis [1] and enantioselective synthesis[2] there has been renewed interest in the synthesis andchemistry of substituted ferrocenes [3]. Ferrocene-basedsystems are widely used as redox sensors. Beer’s ammo-nium ion redox-responsive receptor utilizes a ferrocenemoiety as a simple redox sensor to monitor substratebinding [4]. Ferrocene moieties have been incorporatedinto large proteins to act as redox relays [5]. Amidebond formation between the ferrocene carboxylic acidand the amine-sidechain of lysine is particularly conve-nient for protein labelling. We have recently reported asimple synthetic strategy for the formation of ferro-cenoyl amino acids [6], allowing the N-terminal incor-poration of a redox moiety into larger peptide

assemblies. Ferrocenes exhibit a well-characterized one-electron reversible oxidation wave. Both, ferrocene andferricenium cation are stable under a wide range ofexperimental conditions making ferrocene and itsderivatives extremely useful as an electrochemical probe[7]. Substituents will influence the redox behavior of theferrocene moiety by changing the energy level of theHOMO [8].

In accord with theoretical and experimental results,an electron withdrawing acyl group lowers the energylevel of the highest occupied molecular orbital(HOMO), making it more difficult to oxidize [8]. Forexample, a coplanar arrangement of a phenyl-ring orcarbonyl groups attached to the Cp ring of the fer-rocene moiety, should allow for maximum interactionof the two p-systems and transmittance of ligand influ-ences. Any deviations (e.g. due to sterics) from copla-narity will influence the electronic interaction betweenthe two p-systems.

In this paper, we are particularly interested in under-standing possible electronic factors leading to structuraldistortions of the ferrocenoyl group. In order to investi-

* Corresponding author. Present address: Department of Chem-istry, University of Saskatchewan, 110 Science Place, Saskatoon, SKS7N 5C9, Canada; fax: +1 613 9914278; e-mail:[email protected]

0022-328X/98/$19.00 © 1998 Elsevier Science S.A. All rights reserved.PII S0022-328X(97)00785-7

L. Lin et al. / Journal of Organometallic Chemistry 556 (1998) 11–2012

Table 1Selected structural parameters of ferrocenoyl compounds

B(O–C–E) Cp-bent Cp/COR angleEntry Ref.d(C�O) d(C–C(0))

[10]coplanar1 121.24(11)1.261(15) 1.407 (15)coplanar1 1.232(2) 1.462(2) 123.15(13) this work

[11]0.5/4.1coplanar2 123(1)1.268(10)3.1 (278K)2 1.233(3) 1.468 (3) 123.6(3) [12]

3 1.228(5) 123.5(5) [13][13]4 124.3(3)1.200(4)

2.5/3.2 11.6/5.05 1.243(7) 123.5(5) [14][14]19.82.66a 122.1(6)1.258(5)

6b 1.245(9) 1.476(9) 121.7(6) [15]6c 1.248(8) 1.468(7) 122.1(4) [15]

[15]6d 122.5(12)1.273(12) 1.473(19)126.0(8) 3.65(ecl) 14.39 [16]7 1.244(12) 1.462(14)

[17]15.6/6.08 123.2(5)1.200(5) 1.473(6)1.3 16.79 1.241(11) 1.507(12) 122.3(8) this work

15.7/50.2(!)10 1.222(18) 1.472(19) 120.4(12) [18][17]11 120.9(5)1.232(6) 1.501(7)[19]12 1.235(4) 1.461(5)[20]13 121.5(3)1.219(4) 1.475(4)

14 1.268(3) 1.471(4) 129.6(3) [21]15a 1.222(18) 1.472(19) 120.4(12) [22]

[6]14.116a 122.5(4)1.236(6) 1.499(7)123.6(26) 1.8 5.4 [6]17a 1.21(2) 1.43(2)

[23]6.11.818 120.6(2)1.242(2) 1.501(2)2.8 24.119 1.233(2) 1.504(3) 121.4(2) [23]1.1 17.020 1.238(3) 1.498(3) 121.0(2) B.J. Chapel personal communication0.9 [24]21 1.047(11) 1.481(14)

14.222 1.203(7) 1.460(7) [25]a118.3(5)[26]23 ac 120.5(5)1.210(6) 1.450(7)

benz1. 1.217(6) 1.474(7) 118.1(5)[27]24 1.218 1.471 121.2 1.0

a Two molecules per asymmetric unit.

gate these effects in more detail and to possibly gaininsight in the causes for the variations in electrochemi-cal properties, we carried out electronic structure calcu-lations based on gradient corrected density functionaltheory which has repeatedly shown to provide excep-tionally accurate energetic and structural informationfor transition metal systems [9]. We optimized thegeometry of three representative compounds containingthe ferrocenoyl moiety, ferrocene aldehyde (I), fer-rocene carboxylic acid (II), ferrocene amide (III) andthe parent ferrocene.

Due to our particular interest in ferrocene amides, weare focusing our discussions of electronic and geometricstructures on ferrocene carboxylic acids and amides. Inorder to assess the significance of our calculation, weare presenting them in comparison to experimentalstructural data of known compounds containing theferrocenoyl moiety.

2. Results and discussion

2.1. Geometric structure

2.1.1. Ferrocene carboxylic acidsWe begin our discussion with the parent ferrocene

carboxylic acids. The pertinent structural data for fer-rocene carboxylic acids (entry 1–4)

and some of their metal salts (entry 5, 6a–7) is summa-rized in Table 1. The structure of ferrocene carboxylicacid (1), determined by Cotton et al. [10]a, consists of

L. Lin et al. / Journal of Organometallic Chemistry 556 (1998) 11–20 13

Table 2Optimized geometries of substituted ferrocenes (all distances in A andangles in °)

I IIFerrocene III

1.425–1.442 1.425–1.4441.429–1.4401.432d(C–C)2.059 2.052–2.066 2.035–2.068 2.048–2.072d(C–Fe)

1.2661.2211.220d(C�O)1.485 1.460d(C–C(X)) 1.4601.120 (H) 1.376 (O)d(C–X) 1.463 (N)

122.4120.2 123.5a(O–C–X)178.4177.4a(Cp–Fe–Cp) 180.0 178.44.75.9t(Cp/COX) 4.1

H-bonds between the phosphine oxide and the car-boxylic OH group giving zigzag chains, with a O···O of2.588 (3) A [13].

Ferrocene-1,1%-dicarboxylic acid (2) crystallizes as O–H···O H-bonded dimers in two modifications. The mon-oclinic modification [11] exhibits a disorder showing theH-atom of the acid group equally distributed over twosites between the oxygens atoms of the dimeric unit.The triclinic modification [12] has well ordered posi-tions for the acidic H-atom and well-defined C�O andC–O bonds (d(C�O)=1.233 (3) and d(C–O)=1.300(3) A). In both modifications, the two Cp rings of theferrocene dicarboxylic acid moiety are almost eclipsed(twisted with respect to each other (1.4°) and deviateonly slightly from co-planarity (bent of 1.2°). The O–C–O angle is opened up by ca. 2° as compared to themono-carboxylic acid. The tilt between the Cp-planeand the carboxy-plane (0.5° and 4.1°) allows for effi-cient overlap of the two p-systems.

For comparative purposes, we have carried out a fullgeometry optimization of ferrocene carboxylic acid,(C5H5)Fe(C5H4CO2H) II. The results are summarizedin Table 2 and Fig. 1. On first sight, the overallgeometry is well reproduced. The Fe–C bond distancesare well within the established Cp–Fe distances forother crystallographically determined ferrocene car-boxylic acids (see Table 1 entry 1–4). The carboxy-

hydrogen-bonded dimers with Fe–C bond distances of2.047 (14) A for the substituted Cp ring and a Fe–Cdistance of 2.038 (14) A for the unsubstituted Cp ring[10]. This compares well to simple ferrocene with ad(Fe–C) of 2.05 A [28]. H-bonded dimers having simi-lar structural features are obtained also for systems thatcarry substitutents at the opposite ring. 1-PPh2-1%-COOH(Fc) (3), a centrosymmetric dimer, and 1-P(O)Ph2-1%-COOH(Fc) (4) have their carboxyl groupsin coplanar arrangement with the Cp rings. The Cprings are parallel with a tilt of B3°. 4 exhibits aninteresting solid state arrangement in that it forms

Fig. 1. Optimized geometries of ferrocene (top left), ferrocene aldehyde (I), ferrocene carboxylic acid (II), and ferrocene amide (III). Selected bonddistances and angles are illustrated (also see Table 2).


group is almost co-planar with the Cp-ring to which itis attached (twist of 4.1°). Both Cp rings are nearlyparallel to each other (Cp bent of 2.6°). However, thereare noteworthy differences. The C–C(O) distance wascalculated to be 1.460 A. Interestingly, this is significantlonger than the crystallographically determined bonddistance for 1 (1.407 (15) A), but compares well withthat of 1,1%-ferrocene dicarboxylic acid and other fer-rocene carboxylic acids [10,13,11,12]. We were surprisedto find this discrepancy between our calculated C–C(O)distance and that measured by Cotton [10]a. Hence, wecarried out a careful low temperature single crystalX-ray diffraction study of 1 (for a full analysis, seesupplementary material) [29,30]. Our experimental re-sults for the C–C(O) bond distance of 1.462(2) A is inagreement with our theoretical findings [31].

Ferrocene carboxylates can act as ligands to metalions. Coordination should be manifested in structuralchanges that express a change in electronic structure,such as a change in d(C�O) or change in the tilt of Cpversus carboxy-planes. Structural parameters of someferrocene carboxylates are summarized in Table 1 (en-try 5, 6a–7).

Whereas the dihedral angle between the Cp and thecarboxy planes increase upon coordination of ferrocenecarboxylate to various metals, there is no significantchange in the Cp–C(O) separation. Furthermore, nochanges of the O–C–O angle are observed, nor doesmetal coordination increase the C–O separation. Coor-dination of 1 to Cu2+ causes only a small shift tohigher energy in the Vis absorption from 460 nm in freeferrocene carboxylic acid to 458 nm in 6b, 6c and 455

nm in 5, 6a and 6d [14–16]. Electrochemical studies onthese systems support the UV studies in that the 1-eoxidation wave of the ferrocene/ferrocenium couple isvirtually unaffected by Cu coordination, suggesting thatthe ferrocene orbitals are only slightly disturbed bymetal coordination. This is in accord with theoreticalconsiderations suggesting that a one-electron oxidationwill be Fe-based. Although the highest molecular or-bital (HOMO) will be influenced by the carboxylatesubstituent, more remote changes (such as Cu coordi-nation) are not expected to result in a great variation ofthe HOMO energy. Hence it does not seem surprisingthat no change in redox behavior was observed.

2.1.2. Ferrocene amidesFig. 1 and Table 2 summarize the pertinent struc-

tural features of the optimized structure of ferroceneamide, (C5H5)Fe(C5H4CONH2) III. As expected, bothCp rings are nearly parallel (Cp bent angle 1.6°). Bothamide and Cp planes are close to co-planar with anegligible twist of 4.7°. The C–C(O) distance of III isidentical to that of II. Upon amidation the Fe–C(Cp)

distances and the C�O distances (from 1.221 A in II to1.266 A in III) increase. The C�O bond distances ofstructurally characterized ferrocene amides are between1.20 and 1.26 A (see Table 1) and seem to be related tosteric factors or packing forces rather than electronicfactors. The calculated C(O)–N distance of 1.463 A issignificantly larger than those observed for a series offerrocene amides, which fall in the narrow range of


Fig. 2. Structural representation of FcNHiPr (9). Selected bond lengths and angles: C(1)–C(11)=1.507(12) A, C(11)–O(1)=1.241(11) A,N(1)–C(12)=1.462(12) A, O(1)–C(11)–N(1)=122.3(8)°, C(11)–N(1)–C(12)=121.5(8)°.

1.330(4)–1.368(4) A (see Table 1). However, extremering-strain, such as observed in Hirao’s macrocyclic2-pyridyl-1,1%-ferrocenedicarboximide will lead to anelongation of the C(O)–N bond (1.446 (8) and 1.408 (8)A) [32].

The structural data for some ferrocene amides issummarized in Table 1. In absence of any ring strain, asimple 1,3%-amide such as 8 exhibit a coplanar arrange-ment of the amide groups and the Cp ring [19]. In lightof a dearth of structural information on simple mono-substituted ferrocenoyl amides, we decided to carry outa single X-ray study of FcC(O)NH(iPr) (9) [33,34], firstprepared by Beer et al. [4].

The structure of FcC(O)NH(iPr) is shown in Fig. 2. 9crystallizes in the monoclinic space group P21/c. TheCp rings are parallel to each other (Cp bent angle 1.3°).The Cp–C(O) distance of 1.507(12) A is within therange of other simple ferrocenoyl amides, such as ferro-cenoyl glutamyl dibenzyl ester (16) [6] (1.499(7) A and1.507(7) A) and ferrocenoyl cysteinyl(S-benzyl)methylester (17) [6]

(1.43(2) A and 1.48(2) A) and others (18–20).The amide group C(O)–N is planar and exhibits com-mon C–N and C–O bond distances and angles. It isrotated with respect to the Cp plane by 16.7°. Theisopropyl substituent is rotated by 104.6° with respectto the amide plane, most likely due to a minimizationof steric interaction between the bulky isopropyl sub-stituent on the amide nitrogen and the Cp ring. Similarsteric effects have been observed for benzoyl ferrocene(vide infra). In addition, 9 exhibits extensive hydrogenbonding between the amide N–H and the carbonyloxygen of an adjacent molecules (d(O···N)=2.95 A),forming an infinite chain.

In macrocyclic systems, such as 10, ring strain canlead to extensive distortion from planarity, causing arotation of the Cp and amide planes of 50.2°, withrespect to each other [18]. In Hirao’s imide [3] ferro-cenophane, the Cp/amide rotations are 34.6° and 45.7°


for the two amide groups [32]. In addition, the complexexhibits an unusually large Cp bending angle of 16.4°!In other mono-amides (19–21), only moderate twistsup to 24.1° (for 20) are observed, most likely due tosteric interactions between the Cp ring and the sub-stituent on the amide. Although a coplanar arrange-ment will allow for maximum p-orbital overlapbetween the Cp and amide p-systems, an angle betweenCp and amide planes of 30° will decrease the orbitalinteraction by only ca. 14% and even a twist of 50.2°, asin the case for 9, will still allow for ca. 60% of theorbital interaction (cos u–relationship)! We have esti-mated the barrier for rotation of the amide group aboutthe C–C(O) bond to be about 35–40 kJ mol−1 (esti-

mated as the difference of the total bond energies of IIin Cs and C1 geometry). In general, ring strain causesthe d(C–N) to be elongated with respect to the simpleamide 8, giving it more single bond character.

Experiments show that in the absence of strongintermolecular forces, such as hydrogen bonding (see1,1%-ferrocene dicarboxylic acid, vide supra) [11,12],most of the disubstituted systems crystallize in the 1,3%conformation (e.g. 8). This is not too surprising, sincethis configuration minimizes the steric repulsion be-tween the groups and still allows an eclipsed conforma-tion of the Cp rings. If the two substituents are tiedtogether as part of a macrocyclic system (10 and 11),the ferrocene system adopts a 1,2% conformation due to


Scheme 1. The interaction pattern of the HOMO of ferrocene with the substituent orbitals.

the geometric requirements imposed by the macrocycle.Similar geometries have been observed for ferro-cenophanes [17].

Ferrocenoyl acetylthioureas, generally obtained byreaction of ferrocenoyl chloride with KSCN, followed bythe reaction with primary amines, form an interestingclass of compounds in that they can act as bidentateO,S-ligands to metal centers. Bond distances and anglesare similar to those of simple ferrocenoyl amides (seeTable 1 entry 12–14). Hydrogen abstraction leads to ananionic ligand able to coordinate to transition metal ionsin a bidentate fashion. Expectedly, coordination hasstructural consequences. For example, coordination ofthe ferrocenoyl acetylthiourea ligand to Ni2+ results inelongation of the C�O bond by about 50 pm and ashortening of the C–N bond by about 35 pm. In addition,the O–C–N angle is significantly larger (129.6 (3)°(complexed) as compared to 121.5 (3)° (uncomplexed)).

2.1.3. Ferrocene ketonesFor completion, we wish to include a short section

on ferrocene ketones. Fig. 1 and Table 2 summarize the

important features of the geometry optimization of(C5H5)Fe(C5H4CHO) (I, 21). The overall calculatedstructure is in good agreement with the experiment [24].The two Cp rings are virtually parallel (Cp bent anglecalc. 1.6°; exp. 0.9°) and eclipsed. The –CHO group isalmost coplanar with the Cp ring to which it is attached(twist calc. 5.9°, exp. 6.2°). This arrangement allows forconjugation between the Cp ring and the –CHO group.The Cp–CO bond distance confirms conjugation and isbetween that of a single and double bond (d(C–C) calc.1.485 A; exp. 1.481 (14) A). It is noteworthy that theagreement between experiment and calculation is lessaccurate for the C�O bond length (d(C�O) calc. 1.220A; exp. 1.047 (11) A). The structural features of otherferrocene ketones is summarized in Table 1 (entry21–24).

Similarly to other ferrocenoyl complexes, benzoylferrocene (22) [25]a exhibits signs of double bond local-ization in the substituted Cp ring. This effect is alsoevident from the short Cp–CO distance of 1.467 (3) Aas compared to the Ph–CO of 1.496 (3) A. The amideplane is slightly rotated out of the Cp plane by 14.2°.


Probably due to steric interference between the phenyland the Cp groups, the phenyl substituent is rotated outof the Cp plane by 40.4°. Another structure determina-tion under similar conditions gives similar results hav-ing a a short Cp–CO bond (1.460 (7) A) and anelongated Ph–CO bond (1.491 (8) A) [25]b.

1-Acetyl-1%-benzoyl ferrocene (23) is an interestingcase, since it has both the acetyl and benzoyl groupattached to the Cp rings. Hence it allows a directcomparison of the relative influence of the other sub-stituent with respect to changes in bond distances andangles in the other substituent. The molecule adopts a1,3% configuration, familiar for disubstituted ferrocenes.For the 1-benzoyl substituted Cp ring, the Cp–CO(1.474 (7) A and CO–Ph (1.498 (7) A) bond distancesare almost identical to simple benzoyl ferrocene 1.474(7) A. The benzoyl group is rotated out of the Cp planeby 43.57°. The Cp–CO distance for the 1%-acetyl groupis significantly shorter (1.450 (7) A) indicating it havingmore double bond character (C–C is 1.54 A and C�Cis 1.40 A). The acetyl group is nearly co-planar with theCp ring to which it is attached, allowing for an idealinteraction between the p-systems [26].

In contrast, the acetyl groups in 1,3%-diacetyl-Fc [27]are twisted out of the Cp planes by 11.1 and 8.48°,respectively. The elongation of the Cp–CO distance to1.474 (7) and 1.468 (7) A is most likely due to the effectof the substituent on the other Cp ring. It seems thatthe presence of CO–R group in which R can be in-volved in electron-delocalization leads to a shorteningof the Cp–CO bond on the other ring.

3. Electronic structure

The electronic structure of ferrocene has been exten-sively studied and became a textbook example for thebonding in MCp and MCp2 systems [35]. To interpretthe trends in the structural, spectral, chemical andelectrochemical properties of substituted ferrocenes it isof interest to study the effect of ligand substitution onthe molecular orbitals, especially on the higher lyingoccupied orbitals. The HOMO of ferrocene is an essen-tially non-bonding metal based orbital of e2% symmetry,which correlates to the x2–y2 and xy orbitals of iron in

standard orientation. An iron based z2 orbital is belowthe HOMO of ferrocene followed by two degeneratepairs of bonding orbitals between the p system of theCp ring and the iron d orbitals. These p orbitals play animportant role in substituted ferrocenes since the p

system can extend to the substituents. The ligandsconsidered in this study are all electron withdrawing byaccepting electrons in the unoccupied p* orbitals.

One important effect of ligand substitution is thesplitting between the two components of the degenerateferrocene orbitals. Usually, one of the two componentsof the degenerate orbitals have the proper nodal char-acteristics to interact with the ligand based orbitalswhile the other is non-interacting. We demonstrate thiseffect by considering the two components of theHOMO of ferrocene. These are both metal based or-bitals, but the ligand contribution is different in thesubstituted systems compared to ferrocene. To showthe differences, we consider the ligand contributiononly to these orbitals. In the aldehyde, acid, and amidederivatives the two orbitals which correspond to the 3e2%of ferrocene are separated by 0.046, 0.082 and 0.065 eV,respectively. Other degenerate orbitals which are Cp-based are split even more. For each substituted fer-rocene the ‘a’ component of the 3e2% is stabilized and the‘b’ component is not interacting with the substituentsand becomes the HOMO of the substituted system.Scheme 1 illustrates the general interaction pattern ofthe HOMO of ferrocene with the substituent orbitals.The ‘b’ component of the ferrocene HOMO orbital hasalmost zero contribution from the C1 carbon atom,which makes this orbital less capable to overlap withthe ligand based orbitals. The ‘a’ component (seeScheme 1), however, has a significant contribution fromC1 carbon, hence has some overlap with the orbitals ofthe substituents. The ‘a’ component of ferroceneHOMO is stabilized by a contribution from the car-bonyl oxygen pp orbital, which is a general pattern forall systems.

In addition to the metal based HOMO orbitals, theCp based p orbitals also play a significant role in the inthe electronic structure of substituted ferrocene. Someof the structural differences between the acid and amidederivatives of ferrocene originate from the differences inthe electronic structure of substituents. The p typeorbitals of the substituents can interact with the p

system of the Cp ring on ferrocene. The LUMO of boththe amide and the acid are antibonding p orbitals(Scheme 2) between the C–O and C–X (X=O, N)atom pairs which interact with the occupied p orbitalsof the Cp ring on ferrocene. The LUMO of the amideis lower in energy than that of the acid making theamide slightly more electron withdrawing. The highestoccupied p type orbitals of the amide and the acid alsointeract with the ferrocene. In this respect there is animportant difference between the acid and the amide:Scheme 2. The LUMO of both the amide and acid groups.


the highest p type orbitals of the acid is non-bondingwhile that of the amide is bonding between the car-bonyl C and O atoms. The withdrawal of electrondensity from the highest p type orbital thus leads toweakening and lengthening of the CO bond in theamide more than in the acid.

4. Computational details

The reported calculations were carried out with theAmsterdam Density Functional (ADF) program systemversion 2.0.1 derived from the work of Baerends et al.[36] and developed at the Free University of Amster-dam [37] and at the University of Calgary [38]. Alloptimized geometries calculated in this study are basedon the local density approximation [39] (LDA) aug-mented with gradient corrections to the exchange [39]band correlation [39]c potentials. These optimizationswere carried out by the direct inversion of iterativesubspace for geometry (GDIIS) technique [40] withnatural internal coordinates [41]. We have combinedthe ADF program with the GDIIS program [42] andpreviously implemented the skeletal internal coordi-nates [43]. The internal coordinates were generated bythe INTC program [44] and augmented by hand.

The atomic orbitals on iron were described by anuncontracted triple-z STO basis set [45], while a double-z STO basis set was used for carbon, nitrogen, oxygenand hydrogen; a single-z polarization function was usedon all atoms. The 1s2 configuration on carbon, nitrogenand oxygen as well as the 1s22s22p6 configuration ofiron were assigned to the core and treated by thefrozen-core approximation [38]. A set of auxiliary s, p,d, f, g and h STO functions, centred on all nuclei, wasused in order to fit the molecular density and representthe Coulomb and exchange potentials accurately ineach SCF cycle [46].

The numerical integration accuracy parameter whichapproximately represents the number of significant dig-its for the scalar matrix elements was gradually in-creased to 4.5 until convergence with respect tointegration accuracy was reached. This numerical accu-racy is sufficient to determine energies within a fractionof a kJ mol−1, and bond distances within 0.001 A(which is more accurate than predicted by the numeri-cal integration accuracy parameter).

5. Summary

Despite the multitude of chemical variety of ferro-cenoyl compounds having the chemical formula(C5H5)Fe(C5H4C(O)X), there is surprisingly smallstructural variation with the chemical nature of sub-stituent X. It is obvious from Table 1 that a trend exists

for d(C�O), with d(C�O) of aldehydeBketonesBamidesBacid. The O–C–X (X=O, N, C) angles areless responsive to substituents and seem to be more areflection of packing forces. In addition, d(C(O)–O)Bd(C(O)–N)Bd(C–C). Amides show a narrow range ofC(O)–N distances between 1.330 (4) and 1.368 (4) A.Maximization of electron delocalization involving theC(O)–X (X=O, N, C) and the Cp ring will force theCp ring and the C(O)–X substituent to be co-planar.However, steric interference allows for a significantdeviation from co-planarity. Large substituents on theC(O)–X group can give rise to a significant twist.Likewise ferrocene cyclophanes exhibit larger twiststhan comparable open systems. As pointed out, orbitalinteraction between the p-systems is still operationaleven at large twists of 40–50°. Hence, it can be con-cluded that this twist is most likely a reflection ofcrystal packing forces. All Cp–CO distances show sig-nificant double bond character. As was pointed out byPalenik, the eclisped conformation of the Cp rings ispreferred in the absence of steric contraints and disor-der of the molecules in the crystal state. Using calcula-tions based on density functional theory, we optimizedthe structures of ferrocene aldehyde, ferrocene car-boxylic acid and ferrocene amide. The geometries com-pare favorably with experimental data. Although ourresults suggest that one may predict structural trendssemiquantitatively, our calculations also showed thecomplexity of the problem, which makes it very difficultto predict a priory quantitative trends without detailedcalculations.

Acknowledgements

A. Berces is a Canadian Government Laboratoryvisiting fellow (1995–1997). H.-B. Kraatz is the recipi-ent of an NRC research associateship (1996–1998). Wethank Garry E. Enright for help with the X-ray datacollections and Prof. Peter Pulay for providing theINTC program, used to generate the input for naturalcoordinate optimization.

References

[1] (a) A. Togni, T. Hayashi, Eds., Ferrocenes: Homogeneous Catal-ysis, Organic Synthesis, Material Science, VCH, Weinheim, 1995.(b) M. Sawamura, Y. Ito, Chem. Rev. 92 (1992) 857.

[2] G. Nicolosi, A. Patti, R. Morrone, M. Piatelli, Tetrahedron:Asymmetry 5 (1994) 1639.

[3] (a) T. Sammakia, H.A. Latham, D.R. Schaad, J. Org. Chem. 60(1995) 10–11. (b) M. Tsukazaki, M. Tinkl, A. Roglans, B.J.Chapell, N.J. Taylor, V. Sniekus, J. Am. Chem. Soc. 118 (1996)685–686.

[4] P.D. Beer, D.B. Crowe, M.I. Ogden, M.G.B. Drew, B. MainJ.C.S., Dalton Trans. (1993) 2107–2116.


[5] (a) Y. Degani, A. Heller, J. Phys. Chem. 91 (1987) 1285–1289.(b) A. Heller, Acc. Chem. Res. 23 (1990) 128–134.

[6] H.B. Kraatz, J. Lusztyk, G.D. Enright, Inorg. Chem. 36 (1997)2400–2405.

[7] see for example: (a) L. Zhang, L.A. Godinez, T. Lu, G.W.Gokel, A.E. Kaifer, Angew. Chem. Int. Ed. Engl. 34 (1995)235–237. (b) C.E.D. Chidsey, C.R. Bertozzi, T.M. Putvinski,A.M. Mujsce, J. Am. Chem. Soc. 112 (1990) 4301–4306. (c) A.Badia, R. Carlini, A. Fernandez, F. Battaglini, S.R. Mikkelsen,A.M. English, J. Am. Chem. Soc. 115 (1993) 7053–7060.

[8] J.C. Green, Struct. Bond. 43 (1981) 37–112.[9] (a) T. Ziegler, Chem. Rev. 91 (1991) 651. (b) A. Berces, T.

Ziegler, L. Fan, J. Phys. Chem. 98 (1994) 1584–1595.[10] (a) F.A. Cotton, A.H. Reid, Jr., Acta Cryst. C 41 (1985)

686–688. (b) K. Iwai, M. Katada, I. Motoyama, H. Sano, Bull.Chem. Soc. Jpn. 60 (1987) 1961–1966.

[11] G.J. Palenik, Inorg. Chem. 8 (1969) 2744–2749.[12] F. Takusagawa, T.F. Koetzle, Acta Cryst. B B35 (1979) 2888–

2896.[13] J. Podlaha, P. Stepnicka, J. Ludvik, I. Cisarova, Organometallics

15 (1996) 543–550.[14] A.L. Abuhijleh, C. Woods, J. Chem. Soc. Dalton Trans. (1992)

1249–1252.[15] A.L. Abuhijleh, J. Pollitte, C. Woods, Inorg. Chim. Acta 215

(1994) 131–137.[16] M.R. Churchill, Y.-J. Li, D. Nalewajek, P.M. Schaber, J. Dorf-

man, Inorg. Chem. 24 (1985) 2684–2687.[17] M.C. Grossel, M.R. Goldspink, J.A. Hriljiac, S.C. Weston,

Organometallics 10 (1991) 851–860.[18] C.D. Hall, I.P. Danks, S.C. Nyburg, A.W. Parkins, N.W.

Sharpe, Organometallics 9 (1990) 1602–1607.[19] J.-T. Wang, Y.-F. Yuan, Y.-M. Xu, Y.-W. Zhang, R.-J. Wang,

H.-G. Wang, J. Organomet. Chem. 481 (1994) 211–216.[20] O. Seidelmann, L. Beyer, G. Zdobinsky, R. Kirmse, F. Dietze,

R. Richter, Z. Anorg. Allg. Chem. 622 (1996) 692–700.[21] O. Seidelmann, L. Beyer, R. Richter, Z. Naturrforsch. B 50

(1995) 1684–1684.[22] W. Zhang, F.-Z. Li, Q.-W. Liu, X.-Y. Huang, Jiegou Huaxue

(Chin. J. Struct. Chem.) 14 (1995) 108–112.[23] B.J. Chapell, V. Snieckus, J. Am. Chem. Soc. 118 (1996) 685–

686.[24] M.F. Daniel, A.J. Leadbetter, M.A. Mazid, J.C.S. Faraday

Trans. 2 77 (1981) 1837–1849.[25] (a) I.R. Butler, W.R. Cullen, S.J. Rettig, J. Trotter, Acta Cryst.

C44 (1988) 1666–1667. (b) J.C. Barnes, W. Bell, C. Glidewell,R.A. Howie, J. Organomet. Chem. 385 (1990) 369–378.

[26] (a) G. Calvarin, J. Bouvaist, D. Weigel, C. R. Acad. Sci. Paris268 (1969) 2288–2290. (b) P.G. Calvarin, D. Weigel, Acta Cryst.B27 (1971) 1253–1263.

[27] G. Palenik, Inorg. Chem. 9 (1970) 2429–2430.[28] P. Seiler, J.D. Dunitz, Acta Cryst. B35 (1979) 2020.[29] Crystal data for 1: C11H10FeO2, MW=230.05, orange needles,

crystal size=0.30×0.15×0.15 mm, monoclinic space groupP21/n, a=5.734(2) A, b=12.782(3)A, c=12.550(3) A, b=97.46(3)°, V=912.0(5) A3, Z=4, Dcalc.=1.595 g cm−1, l(Mo–Ka)=0.71073 A (graphite monochromated), F(000)=452,m=1.613 cm−1, R=0.0226 and wR=0.0574 for I\2s(I),R=0.0269 and wR=0.0600 for all data.

[30] Data collection was carried out on a Siemens SMART diffrac-tometer equipped with a CCD collector. The structure wassolved by direct methods and refined on F2 data using theSiemens SHELXTL program system. All non hydrogen atoms wererefined with anisotropic temperature factors. Hydrogen atomswere located in the density map and positions fully refined usingisotropic temperature factors. The refinement converged at R=

0.0226, wR=0.0574 (for all observed data I\2s(I).[31] According to our experimental findings, the structure of fer-

rocene carboxylic acid consists of hydrogen bonded acid dimerswith d(O···O)=2.656 A, which is in good agreement with Cot-ton’s structure (see ref. [11]a). In addition, we find a weakinteraction of between the oxygen of the carboxylate with a C–Hbond of the Cp of d(O···H)=2.534 A. This distance is clearlywithin the sum of the van der Waals radii and indicates a weakinteraction.

[32] T. Moriuchi, I. Ikeda, T. Hirao, Organometallics 14 (1995)3578–3580.

[33] Crystal data for FcC(O)NH(iPr): C14H17FeNO, MW=271.14,thin yellow needles, crystal size=0.45×0.05×0.02 mm, mono-clinic space group P21/c, a=13.1973(10) A, b=9.8447(8) A,c=10.1206(8) A, b=109.2590(10)°, V=1241.3(2) A3, Z=4,Dcalc.=1.451 g cm−1, l(Mo–Ka)=0.71073 A (graphitemonochromated), F(000)=568, m=1.197 cm−1, R=0.0979and wR=0.2300 for I\2s(I) (R=0.1474 and wR=0.2724 forall data). Due to the small crystal size only one quarter of allreflections were observed.

[34] Data collection was carried out on a Siemens SMART diffrac-tometer equipped with a CCD collector. The structure wassolved by direct methods and refined on F2 using the SiemensSHELXTL program suite. All non-hydrogen atoms are refinedanisotropically. Hydrogen atoms are placed in calculated posi-tions (see supplementary material).

[35] T.A. Albright, J.K. Burdett, M.-H. Whangbo, Orbital Interac-tions in Chemistry, John Wiley and Sons, New York, 1985.

[36] E.J. Baerends, D.E. Ellis, P. Ros, Chem. Phys. 2 (1973) 41.[37] (a) W. Ravenek, in: H.J.J. te Riele, Th. J. Dekker, H.A. van de

Vorst (Eds.), Algorithms and Applications on Vector and Paral-lel Computers, Elsevier, Amsterdam, 1987. (b) P.M. Boerrigter,G. te Velde, E.J. Baerends, Int. J. Quantum Chem. 33 (1988) 87.(c) G. te Velde, Baerends, E.J., J. Comp. Phys. 99 (1992) 84.

[38] (a) L. Fan, T. Ziegler J. Chem. Phys. 94 (1991) 6057. (b) L. Fan,T. Ziegler, J. Chem. Phys. 95 (1991) 7401. (c) L. Fan, L.Versluis, T. Ziegler, E.J. Baerends, W. Ravenek, Int. J. QuantumChem. S22 (1988) 173. (d) L. Fan, T. Ziegler, J. Chem. Phys. 96(1992) 9005. (e) L. Fan, T. Ziegler, J. Phys. Chem. 96 (1992)6937.

[39] (a) S.H. Vosko, L. Wilk, M. Nusair, Can. J. Phys. 58 (1980)1200. (b) A.D. Becke, Phys. Rev. A 38 (1988) 2398. (c) J.P.Perdew, Phys. Rev. B 33 (1986) 8822. (d) J.P. Perdew, Phys. Rev.B34 (1986) 7046.

[40] (a) P. Csaszar, P. Pulay, J. Mol. Struct. 114 (1984) 31. (b) P.Pulay, Chem. Phys. Lett. 73 (1980) 393. (c). P. Pulay, J. Comput.Chem. 3 (1982) 556.

[41] (a) G. Fogarasi, X. Zhou, P.W. Taylor, P. Pulay, J. Am. Chem.Soc. 114 (1992) 8191. (b) P. Pulay, G. Fogarasi, F. Pang, J.E.Boggs, J. Am. Chem. Soc. 101 (1979) 2550.

[42] Programmed by A.G. Csaszar, P.G. Szalay, Eotvos University,Budapest, Hungary, 1984.

[43] (a) A. Berces, J. Comp. Chem. 18 (1997) 45–55. (b) A. Berces, T.Ziegler, J. Phys. Chem. 98 (1994) 13233. (c) A. Berces, T. Ziegler,L. Fan, J. Phys. Chem. 98 (1994) 1584.

[44] INTC program to generate natural internal coordinates, P.Pulay, G. Fogarasi, 1992, Univerity of Arkansas, Fayetteville.

[45] (a) G.J. Snijders, E.J. Baerends, P. Vernooijs, At. Nucl. Data.Tables 26 (1982) 483. (b) P. Vernooijs, G.J. Snijders, E.J.Baerends, Slater Type Basis Functions for the whole PeriodicSystem, Internal report, Free University of Amsterdam, TheNetherlands, 1981.

[46] J. Krijn, E.J. Baerends, Fit functions in the HFS-method, Inter-nal Report (in Dutch), Free University of Amsterdam, TheNetherlands, 1984.

.

ferrocenic acid derivatives: towards rationalizing changes in the electronic and geometric...

Documents